RAID is an acronym for Redundant Array of Inexpensive Disks. In fact, RAID technology is a way to combine several disks to a single storage. Disks that form an array are called "member disks".
Besides the fact that RAID technology is designed to combine the disks, it also can provide performance improvement and fault tolerance due to data redundancy, sometimes at a cost of leaving some disk space unusable. For a detailed information for your specific case, use the free online RAID calculator.
There are various types of RAID (also called RAID levels) which differ in the data placement patterns and the degree of redundancy. In practice, RAID can be implemented either based on dedicated hardware (hardware RAID) or by means of operating system (software RAID).
Now, a software RAID is the cheapest RAID implementation. Many operating systems are capable of creating some RAID types. For example, Windows home editions are able to create only RAID0, Windows server editions allow user to create RAID0, RAID1, and RAID 5. In Linux software RAIDs are implemented based on mdraid or LVM drivers.
Keep in mind, that a software RAID uses host system resources, e.g. CPU, to operate. However, the negative impact on performance is in most cases negligible (see this blog post for a sample calculation).
Software RAIDs have some limitations. It is not possible to boot from a software RAID 0 or RAID 5, and they do not officially support hotswap. However, software RAIDs provide the maximum possible compatibility - just put the array into the different computer running a similar operating system, and the array is ready to use.
Hardware RAID is more expensive type of RAID implementation. It can be created based on the chip integrated into the motherboard (relatively inexpensive) or using standalone feature-rich controller with own CPU and a number of other options (the most expensive RAID implementation).
Continue to RAID levels.